The ability to attribute mental states such as perception, attention, desire, or belief to ourselves and others is critical to humans' social and cognitive competence. With this ability we can communicate referentially, predict and explain each others' behaviors, and manipulate both our own and others' mental states for the purposes of complex problem-solving and learning, not to mention deception. When and how this ability becomes available to children is therefore of particular theoretical importance. The current aim is to examine several unresolved questions in this regard. When can we justifiably claim that infants attribute mental states to others? When they do, who do they attribute mental states to and on the basis of what evidence? The answer to these questions will in turn provide insight into the representational systems underlying mentalistic reasoning. The research proposed will begin with the findings of Johnson, Slaughter, and Carey (1998) in which infants were found to follow the "gaze" of novel objects that embodied characteristics of mentalistic beings, in particular the ability to interact contingently with other agents. This result has been interpreted as a reflection of the infant's attribution of attention and/or perception to the object, i.e., infants follow the object's "gaze" in order to find what it is "looking at." Two complementary series of studies are proposed to examine this interpretation. The first series will attempt to more fully characterize the parameters and representation of contingent interaction as it functions in eliciting gaze-following both in infants and adults. The second series of studies will examine the meaning that infants assign to objects identified via contingency information by examining its effectiveness in eliciting putatively mentalistic attributions across a wide variety of behavioral contexts including imitation, the elicitation of communicative gestures, and the violation of expectations based on mentalistic interpretations.